About NOSE

The same name kept showing up as a different jar.

NOSE reads the terpene panel on a lab report and turns it into a flavor fingerprint, then compares it to jars you already liked. That is the entire product. It will not tell you how anything will make you feel — nobody can tell you that from a terpene panel, and we would rather say so than sell you a guess.

Why we built it

We had twenty years of notes and nowhere to put them.

Buying by name stopped working

Between us, twenty years of buying flower. Most of that time we bought by name, because a name was the only handle anyone ever gave us.

It worked for a while. Then the market got bigger and the names stopped meaning much. The same name on two shelves became two different plants — different grower, different genetics, different cure. Even the same grower’s next run comes back smelling different from the last one. There is no registry, no standard, nothing holding a name to a plant.

Our noses still worked fine

We knew within about a second of the lid coming off whether a jar was worth it. That part never got worse.

We just had no way to write it down, no way to ask for it again, and no way to explain it across a counter to someone who wanted to answer a completely different question.

What changed

The information was already printed on the jar.

Lab reports

In legal markets every batch ships with a COA — a certificate of analysis listing the terpene percentages for that exact jar. The thing we had been guessing at for two decades was sitting on a document almost nobody reads.

So we started reading them. Then saving them. Then comparing them.

The correction everything rests on

The thing we got wrong first is the thing NOSE is now built on: you cannot compare raw percentages. A jar at 3.2% total terpenes and a jar at 1.1% can be the same aroma at two different volumes.

Raw numbers measure loudness. To compare character, you convert each report into shares of its own total — what fraction of this jar’s aroma is limonene, myrcene, caryophyllene — and compare the shape.

Normalized to share-of-total Loudness kept separate

That one correction is the difference between a tool that works and a spreadsheet that quietly lies to you. Everything else here is downstream of it.

Standing rules

What we won’t do.

We won’t tell you how it will feel

Not relaxing, not energizing, not good for focus. The science connecting terpenes to aroma is solid. The science connecting them to effects, at the concentrations actually present in flower, is not there yet. Plenty of people will tell you otherwise with total confidence. We would rather be the site that still holds up in five years.

We won’t invent numbers

NOSE is new. It has no user count worth bragging about, no five-star average, no lab of its own, no white coats. When you see a trust claim on this site it is something structurally true today — third-party lab data, a published method, nothing stored or sold — not a statistic we made up to look established.

We’ll say when we’re unsure

The core idea behind NOSE — that two jars with a similar terpene shape will be similarly enjoyable to the same person — is a reasonable hypothesis. It is not a proven one.

Stated plainly on the methodology page, not buried
Thumbs up and thumbs down on every single match
Your palate stays in your browser unless you export it

We won’t trust the label over the report

Strain names drift, batches drift, and nothing enforces either. The lab report for the specific jar in front of you is the only thing NOSE treats as real.

Who this is for

The most useful thing you can do is tell us we’re wrong.

We built this for the version of ourselves standing in a shop, trying to describe a jar from four months ago, and giving up. If that’s you, there’s nothing to buy — run your own lab reports through the matcher and hit thumbs-down when we miss. A bad call tells us something no marketing page ever will, and we read them.

That’s the whole arrangement: you know what your nose likes, we know what the numbers say, and the gap between those two is the only interesting problem here.

Build a palate from a jar you loved Read the method first